Technology

XDN™ was built on a Next-Generation, fully generic Content Delivery technology stack:

Before being used as the brains and brawn behind the XDN™ platform, these have been originally used and stress-tested to build fully functional private content delivery networks.

XDN™ showcases a set of products that have proven track record of leveraging highly distributed infrastructures and managing them as one unified and consistent Content Delivery Network.

Each component of the XDN™ was designed from the ground up, integrating with the most widespread existing Content Delivery software out there, all following this set of principles:

  1. Highly distributed
  2. Highly resilient
  3. Embraces graceful failure as a design principle
  4. Deeply tunable and elastic
  5. Compatible with existing Content Delivery Networks and Server technologies
  6. API-centric, making it trivial to integrate with your existing backend

CrowdDirector™

CrowdDirector™ is our universal purpose-built content load balancer In The Sky.

It is the foundation of the XDN™ technology stack.

Load balancing as a service

We have spun CrowdDirector redirector nodes all across the globe and packaged them with the lightest footprint possible to be able to spin them in multiple additional geographies in seconds as the need arises.
CrowdDirector™ is the service part of the XDN™ formula and is fully operated by our Operations engineers.

Criteria-rich load balancing

CrowdDirector™ was built in a way that it can generically route traffic between virtually any Content Delivery infrastructure, whether using HTTP 302 redirects or DNS redirects.
Routing rules available for CrowdDirector™ include:

  • Time of day, schedules
  • ASNs, IP Addresses
  • Location (including metros)
  • Regex based URL inspection
  • many other HTTP 302 specific hooks…

Stateful load balancing

CrowdDirector™ integrates tightly with the CrowdCache™ edge proxy, offering health, performance and capacity hooks to tie all the Infrastructure Providers’ XDN™ resources together in a unified, self-healing delivery network.

CrowdCache™ edge nodes enter and exit the pools of routing resources depending on their health, performance and headroom.
When used with 3rd party delivery resources, the health-check system will ensure no content request gets routed to failing content delivery assets.


Sign Up!

CrowdDirector™ tech specs at a glance:

Globally anycast DNS/HTTP redirectors (computing of inbound requests)

  • Global and Regional DNS caching layers
  • Seamless failover to second closest redirector
  • Redirectors in  Amsterdam, Tokyo, San Jose, London, Chicago, Sydney, Miami, Singapore

Zero Software/Hardware needed: Cloud Load Balancer

  • Entirely actionable/configurable via Web Interface
  • Integration API for your backend
  • Integrates with Any webserver, CDN or CrowdCache™ fleet

Dual HTTP 302 redirects and DNS load-balancer

  • DNS and HTTP criteria: ASN, IP(CIDR), Geo, Date/Timerange, daily schedule, monthly schedule
  • HTTP criterion: URL pattern matching
  • Primary and Failover delivery actions
  • Even, Weighted and Auto-Weighted delivery strategies
  • Can load-balance any protocol using DNS rules
  • Rules Cloning/inheriting capability

Highly configurable health-checking backend

  • HTTP,DNS,TCP,ICMP
  • Configure polling intervals, geography, up & down thresholds
  • Extensive notification system: email, SMS, HTTP POST

Powerful analytics:

  • Hits by delivery resources, Rules
  • Geo, ASN, Browsers breakdown
  • All graphs exportable through CSVs
Integration API to manipulate Rule Sets and integrate into white labeling

CrowdCache™

CrowdCache™ brings to the table the integrated control plane that every hardware, off-the-shelf and open-source caching solution lacks.

…with the lightest weight and footprint possible.


Light-weight reverse proxy-cache in a box.

CrowdCache™ is a full-blown Edge Proxy Caching server. With a file size of merely 5MB, this small footprint Java binary will turn any server into a powerful edge proxy-caching server.

A proxy with a unified control plane

Not only will this server literally take seconds to deploy, but it will also register instantly with the XDN™ crowd-sourced delivery network after booting.

When booting, a CrowdCache™ node will establish persistent connections with the health-checking backend, origin and edge-mapping configuration repository and CrowdDirector Rules engine. Any change in the aforementioned will be applied to the overall fleet of CrowdCaches™ in near real time.

Feature-rich edge caching

CrowdCache™ natively includes all features found in legacy CDNs: HTTP Pseudo-Streaming for FLV and MP4, secure tokens, short-lived URLs, rate-throttling by content type, automatic based on meta-data…
Running a unified CrowdCache™ fleet of servers also brings a significant byproduct: all features added to the software in the future will be loaded remotely, without service interruption, rendering the new feature rollouts completely transparent for both Content Provider customers and Infrastructure Provider partners.


Sign Up!

CrowdCache™ service specs at a glance:

lightweight: <5MB binary
small footprint: sealed Jar file, no external need than Java 1.6 needed.
JAR or ISO deployment
Massively distributed, operated through CrowdDirector Web GUI
  • Macro-management through grouping
  • Micro-management through CLI
Seamless connection to CrowdDirector™ health-checking and performance backend
Seamless integration to the CrowdDirector™ analytics backplane:
  • Redirector view
  • Content view
  • Infrastructure view
Deeply configurable cache behavior:
  • Proxy-thru and Redirect modes
  • On cache miss action, on Cache feed action, on Proxy-thru action
    • Dismiss Query Args
    • Replace Query Args
    • Force Query Args
  • API controlled cache flush
  • Ignore caching for specific MIME Types
Tokenized, short-lived URLs
3rd universal token adaptor
HTTP Video Seeking:
  • FLV (bytes,time)
  • MP4 (time)
Rate shaping
  • Auto, based on metadata
  • Global
  • By MIME type
Raw Logs
  • Origin
  • Edge
  • Content

CrowdDirector™, CrowdCache™ and statistics API

“One API to rule them all…”

Service point oriented platform

Every component designed on the CrowdDirector™ and CrowdCache™ platforms interact and are configured through an API. Whether the desired action is to withdraw info on the objects configured in the XDN™ world, to configure them, edit them or delete them, everything equals to calling a service point.
Since we internally use API to interconnect all elements of our platforms, it is only natural that we make some of them public for our customers to:

  • Integrate with their own backend and delivery strategy
  • Build Apps that unveil new possibilities
  • Automate some of the most common manual tasks they would initially perform through the Web UI of CrowdDirector™

Available API service points

At current, you can, amongst others:

  • get/clone/edit/delete CrowdDirector Rule Sets
  • get/set/edit/delete CrowdCache Origin-to-Edge mappings
  • get traffic statistics: hits, traffic

Let us know of any specific need to expand the API scope, we will likely only need to make public some endpoints we already use internally within the platform.

Forgot?